Granted Patent B2

US 10,922,314 · App. 15/563,124 · Granted Feb 16, 2021

Inventors: QiFan Chen (Austin, TX); Choudur Lakshminarayan (Austin, TX)

Assignee: MICRO FOCUS LLC

G06F16/24545G06F16/2358G06F16/2379G06F16/2455G06F17/18

Abstract

Claims (45)

1. A non-transitory computer readable storage medium storing instructions that, when executed by a computer, cause the computer to:

access a plurality of datasets sampled from a column of a database at different times;

compare a first skew associated with a first dataset of the plurality of datasets to a second skew associated with a second dataset of the plurality of datasets, based on a comparison of a first value for a statistic derived from the first dataset with a second value for the statistic derived from the second dataset, wherein the first skew or the second skew is a degree of change in a respective dataset;

based on the comparison of the first skew to the second skew, selectively combine a subset of data from the first dataset and a subset of data from the second dataset to provide a combined subset of data;

use the combined subset of data to determine a third value for the statistic;

based on the comparison of the first skew to the second skew, use the second value or the third value for the statistic to estimate a count of unique values in the column of the database;

receive a query to retrieve data from the database;

generate an execution plan for the query based on the count of unique values in the column; and

execute the query based on the execution plan to provide query results.

2. The non-transitory computer readable storage medium of claim 1 , wherein the instructions, when executed by the computer, further cause the computer to apply a linear weighted combination estimator to estimate the count of unique values using the second value or the third value for the statistic.

3. The non-transitory computer readable storage medium of claim 1 , wherein the statistic comprises a frequency of unique values appearing in the column.

4. The non-transitory computer readable storage medium of claim 1 , wherein, to generate the execution plan for the query, the instructions further cause the computer to perform selecting operators and selecting an operator execution order for the query based on the count of unique values in the column.

5. The non-transitory computer readable storage medium of claim 1 , wherein the instructions, when executed by the computer, further cause the computer to determine a probability of rows of the database being sampled more than once among at least two datasets of the plurality of datasets and further base selective combination of the subsets of data from the first and second datasets based on the probability.

6. A system comprising:

at least one hardware processor; and

a memory storing instructions that when executed cause the at least one hardware processor to:

access a plurality of datasets sampled from a column of a database, each dataset of the plurality of datasets being associated with a different sampling period of a plurality of sampling periods;

compare a first degree of change associated with a first dataset of the plurality of datasets to a second degree of change associated with a second dataset of the plurality of datasets based on a comparison of a first value for a statistic derived from the first dataset with a second value for the statistic derived from the second dataset;

based on the comparison of the first decree of change to the second degree of change, selectively combine a subset of data from the first dataset and a subset of data from the second dataset to provide a combined subset of data;

use the combined subset of data to determine a third value for the statistic;

based on the comparison, use the second value or the third value for the statistic to estimate a count of unique values in the column of the database;

receive a query to retrieve data from the database;

generate an execution plan for the query based on the count of unique values in the column; and

execute the query based on the execution plan to provide query results.

7. The system of claim 6 , wherein: the statistic comprises a frequency of unique values appearing in the column; and the instructions, when executed, cause the at least one hardware processor to: determine a first value for the frequency of unique values appearing in the column based on the first dataset, and determine a second value for the frequency of unique values appearing in the column based on the second dataset;

determine a difference between the first and second values;

determine a standard deviation for the frequency of unique values appearing in the column based on the difference between the first and second values; and determine degree of change based on the first value, the second value and the standard deviation.

8. The system of claim 6 , wherein the instructions, when executed, cause the at least one hardware processor to:

determine a probability of data entries of the database being sampled more than once in multiple datasets of the plurality of datasets, and

in response to the probability, discard the combined subset of data and generate the third value for the statistic based on the subset of data from the second dataset which is associated with a most recent sampling period.

9. The system of claim 6 , wherein the instructions, when executed, cause the at least one hardware processor to apply a linear weighted combination estimator to estimate the count of unique values in the column using the second value or the third value for the statistic.

10. The system of claim 6 , wherein the statistic indicates a frequency of unique values appearing in the column.

11. A method comprising:

accessing, by a processor, a plurality of datasets sampled from a column of a database at different times;

comparing, by the processor, a first skew associated with a first dataset of the plurality of datasets to a second skew associated with a second dataset of the plurality of datasets, based on a comparison of a first value for a statistic derived from the first dataset with a second value for the statistic derived from the second dataset, wherein the first skew or the second skew is a degree of change in a respective dataset;

based on the comparison of the first skew to the second skew, selectively combining, by the processor, subsets of data from the first dataset and from the second dataset of the plurality of datasets to provide a combined subset of data;

using, by the processor, the combined subset of data to determine a third value for the statistic;

based on the comparison of the first skew to the second skew, using, by the processor, the second value or the third value for the statistic to estimate a count of unique values associated with the column;

receiving a query to retrieve data from the database;

generating, by the processor, an execution plan for the query based on the count of unique values in the column; and

executing, by the processor, the query based on the execution plan to provide query results.

12. The method of claim 11 , further comprising applying a linear weighted combination estimator to estimate the count of unique values in the column using the second value or the third value for the statistic.

13. The method of claim 11 , wherein the statistic indicates a frequency of unique values appearing in the column.

14. The method of claim 11 , wherein generating the execution plan for the query comprises selecting operators and selecting an operator execution order for the query based on the count of unique values in the column.

15. The method of claim 11 , further comprising determining a probability of rows of the database being sampled more than once among at least two datasets of the plurality of datasets and further selectively combining the subset of data from the first dataset and the subset of data from the second dataset based on the probability.

Assignments (9)

Continuity (1)